R help - Aug 2011 - How to find the accuracy of the predicted glm model with family = binomial (link = logit)

Hi All,

When modeling with glm and family = binomial (link = logit) and response values
of 0 and 1, I get the predicted probabilities of assigning to my class one, then
I would like to compare it with my vector y which does have the original labels.
How should I change the  probabilities into values of zero and 1 and then
compare it with my vector y to find out about the accuracy of my prediction?

Thanks,
Andra

Andra,

Two common methods are the AUC (area under the curve) and the Log Likelihood.

The ROCR package has several nice functions for evaluating the accuracy of your
model


--
Noah Silverman
UCLA Department of Statistics
8117 Math Sciences Building
Los Angeles, CA 90095

On Aug 25, 2011, at 9:58 PM, Andra Isan wrote:
> Hi All,
> 
> When modeling with glm and family = binomial (link = logit) and response
values of 0 and 1, I get the predicted probabilities of assigning to my class
one, then I would like to compare it with my vector y which does have the
original labels. How should I change the  probabilities into values of zero and
1 and then compare it with my vector y to find out about the accuracy of my
prediction?
> 
> Thanks,
> Andra
> 
> ______________________________________________
> R-help@r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide
http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.

	[[alternative HTML version deleted]]

Andra,

You can transfer the probabilities into class membership by setting up
a cut-off value, usually 0.5.

set.seed(3)
x<-rnorm(100)
y<-rbinom(100,prob=1/(1+exp(x)),size=1)
model<-glm(y~x,family=binomial(link=logit))
pred<-ifelse(predict(model,type='response')>0.5,1,0)
table(pred,y)


Weidong Gu

On Fri, Aug 26, 2011 at 12:58 AM, Andra Isan <andra_isan at yahoo.com>
wrote:
> Hi All,
>
> When modeling with glm and family = binomial (link = logit) and response
values of 0 and 1, I get the predicted probabilities of assigning to my class
one, then I would like to compare it with my vector y which does have the
original labels. How should I change the ?probabilities into values of zero and
1 and then compare it with my vector y to find out about the accuracy of my
prediction?
>
> Thanks,
> Andra
>
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide
http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
>